Should C assertion failures have the "type-crash" label?

[alex_Gaynor]

As part of triaging bugs found by the fuzzers, I’ve filed a number that involve C assertion failures. For example:

• Objects/call.c:342: PyObject *_PyObject_Call(PyThreadState *, PyObject *, PyObject *, PyObject *): Assertion `!_PyErr_Occurred(tstate)' failed. · Issue #113602 · python/cpython · GitHub
• Python/flowgraph.c:483: _Bool no_redundant_jumps(cfg_builder *): Assertion `last->i_target != b->b_next' failed · Issue #113603 · python/cpython · GitHub

For bugs like these, should I be attaching the “type-crash” label on Github, or is that only for crashes impacting non-debug builds?

[pitrou (Antoine Pitrou)]

If a C assertion would fail in debug mode, chances are that you would get a crash later if in release mode. My vote would therefore be on attaching the “type-crash” label.

[gpshead (Gregory P. Smith)]

Agreed, they are crashes.

[alex_Gaynor]

Sounds good. Thanks all.

Happy New Year!